Case studies & research

Real questions. Evidence you can work with.

See how we turn an AI-search question into a useful diagnosis, a specific action and a testable next step. Selected research spans local commerce, software and our own Web3 GEO site.

Six practical studies

From source clarity to AI answers.

Three independent public website analyses, two anonymized diagnostics and one own-site implementation case. Each records its scope, observation date and current stage. Public analyses do not imply a client relationship or endorsement.

PUBLIC WEBSITE REVIEWRESEARCH NOTE

Make the buying facts agree.

1
pricing page reviewed: 2
conflicting order limits: 0
AI-answer tests in this study

Document the source → confirm the facts → validate the change

Summary of recorded observations · 2026-10-24

Pricing clarity · Commerce software

RateUp: one Free plan, two order limits.

A merchant comparing a free shop plan encountered conflicting order limits on the same pricing page. We documented the conflict and drafted two corrections for the product owner to choose between.

Scope: 1 pricing page · table + FAQ
Platforms: Public website review

Independent public analysisRead the case →

PUBLIC WEBSITE REVIEWRESEARCH NOTE

Make the buying facts agree.

2
visible product families: 1
family in reviewed offer list: 0
AI-answer tests in this study

Document the source → confirm the facts → validate the change

Summary of recorded observations · 2026-10-24

Product clarity · Business software

Cleomitra: two products need two buying answers.

The pricing page showed both Engage and Service OS, while the reviewed software-offer data represented Engage only. We mapped a minimal way to bring the product explanation and offer data into agreement.

Scope: 1 pricing page · visible plans + offer data
Platforms: Public website + structured data

Independent public analysisRead the case →

PUBLIC WEBSITE REVIEWRESEARCH NOTE

Make the buying facts agree.

1
service page reviewed: 1
offer destination to align: 0
AI-answer tests in this study

Document the source → confirm the facts → validate the change

Summary of recorded observations · 2026-10-24

Service identity · Web3 security

Three Sigma: align the offer with the audit scope.

An economic-audit service and its structured offer pointed to different audit scopes. We documented the exact mismatch and proposed a small correction that preserves the rest of the service data.

Scope: 1 service page · identity + offer URL
Platforms: Public website + structured data

Independent public analysisRead the case →

ANSWER ACCURACYRESEARCH NOTE

Same need. Different conditions.

6
buyer questions: 12
captured AI answers: 1
test round

Confirm the policy → reconcile pages → retest

Summary of recorded observations · 2026-10-12

Answer accuracy · Local commerce

One delivery question. Two different AI answers.

A delivery-policy review found conflicting AI answers to a purchase-critical question. The useful next step was to clarify the service facts before writing more content.

Scope: 6 questions · 1 round
Platforms: ChatGPT + Gemini

Anonymized researchRead the case →

BUYER DISCOVERYRESEARCH NOTE

Relevant on-site. Absent in this snapshot.

1
buyer scenario: 2
captured AI answers: 0
target mentions in those answers

Verify capability → clarify the page → repeat

Summary of recorded observations · 2026-10-13

Buyer discovery · Appointment software

A relevant product missed the buyer shortlist.

The website described the requested appointment workflow, but two captured AI answers named other tools. We turned that small observation into a focused product-page verification plan.

Scope: 1 buyer question · 1 round
Platforms: ChatGPT + Gemini

Anonymized researchRead the case →

MEASUREMENT QUALITYRESEARCH NOTE

Keep the unknowns visible.

5
historical web-search queries: 0/5
molthub mentions in that observation: 403
audit access response on 3 Oct

Observe → implement → verify the outcome

Summary of recorded observations · 2026-10-03

Measurement · molthub self-audit

We put our own GEO measurement under review.

Our own site became the test case: record the baseline, keep access failures visible and repair the scanner so that an unreadable robots policy cannot count as a pass.

Scope: 5 historical search queries
Platforms: Web search + access checks

Own-site case studyRead the case →

How to read these cases

The external studies show diagnosis and recommended actions; the own-site case includes implemented fixes. Follow-up visibility and sales results are reported only when measured.

Start with your buying question

What should AI get right about your business?

Run a Free Scan